The definition of open sets I'm working with (from the Napkin) is:
A set $U \subseteq M $ is open if for all $p \in U$, some $r$-neighborhood ($r > 0$) exists in $U$.
The question I'm trying to solve is this:
Are $\mathbb{Q}$ and $\mathbb{N}$ homeomorphic subspaces in $\mathbb{R}$?
My response was no, because any subset of $\mathbb{N}$ is closed, since any sequence that converges using points in $\mathbb{N}$ converges to something in $\mathbb{N}$, whereas there are subsets of $\mathbb{Q}$ that aren't closed - i.e. $3, 3.1, 3.14, 3.141, 3.1415 ... $ that converge to $\pi \in \mathbb{Q'}$.
But the Napkin's solution I don't understand - it simply says that no singleton set is open in $\mathbb{Q}$, but all singleton sets are open in $\mathbb{N}$. I don't understand how I can apply this definition to get a singleton set in $\mathbb{N}$ that's open. Furthermore, I don't even understand what $U$ is in that definition, is it the singleton set or $\mathbb{Q}$ in this context?
Can you explain how to use the above definition to get the result the Napkin did?
Denote by $B(x,r)$ the open ball centered on $x$ with radius $r >0$. $B(x,r)$ is an open subset of the reals... good to know that an open ball is open!
Consider $n \in \mathbb N$. Then $B(n, 1/2) \cap \mathbb N = \{n\}$ and therefore any singleton of $\mathbb N$ is an open subset.
Now if $\mathbb Q$ and $\mathbb N$ would be homeomorphic for the usual topology the image of any open subset of $\mathbb N$ would be isomorphic to its image under the homeomorphism. But that can’t be as it exists open singleton in $\mathbb N$ while any open subset of $\mathbb Q$ contains an infinite number of rationals.